Back to guides

Empryo Integration

Set up Empryo, the graph-powered AI coding agent, with LLM Gateway. One browser login connects the terminal UI, desktop app, and headless CLI to every model on the gateway.

Empryo is a graph-powered AI coding agent from empryo.com — it parses your codebase into a code-intelligence graph and edits symbols, not strings. It ships as a terminal UI, a desktop app, and a headless CLI on macOS, Linux, and Windows, and LLM Gateway is built in: one browser login connects all ~190 text models on the gateway, and every request is tagged with x-source: empryo so usage is attributed correctly in your dashboard and on your DevPass plan.

On a DevPass plan? Log in with /login llmgateway-devpass and use canonical model IDs (claude-opus-5, not anthropic/claude-opus-5) — provider-pinned routing is not available on coding plans; the gateway picks the provider for you.

Install

macOS / Linux:

1curl -fsSL https://empryo.com/install.sh | bash

Windows (PowerShell):

1irm https://empryo.com/install.ps1 | iex

Each script downloads the release's SHA256SUMS.txt, checks its RSA-4096 signature against a key pinned in the script, and verifies the binary's checksum before installing it — the one-liner itself is a bootstrapper, so if you'd rather not pipe it into your shell, download it first and read it, or grab a binary and its checksum from the download page and verify by hand.

Setup

Step 1: Launch Empryo

Start Empryo in any project directory:

1empryo

Step 2: Log in with your browser

Type /login. Empryo opens llmgateway.io/connect/cli and starts a local loopback server — approve the connection in the browser and the minted API key is delivered straight back to Empryo. Nothing to copy or paste. On a DevPass coding subscription, run /login llmgateway-devpass instead.

Prefer to paste a key? Use /keys inside the TUI, run empryo --set-key llmgateway <key> from a shell, or set the LLM_GATEWAY_API_KEY environment variable.

Step 3: Pick a model

Open the model picker with /models. LLM Gateway's models are grouped by upstream provider, with live pricing, context windows, and output limits pulled from the gateway's own catalog — what you see is what you're billed.

Step 4: Start coding

Every prompt and its cost appears in your dashboard, attributed to Empryo. On a DevPass key, Empryo also renders your plan meters — monthly credits and the weekly premium window — as a footer badge in the TUI and in the desktop usage drawer, next to Empryo's own local usage ledger (/usage).

One connection, three surfaces

The gateway connection is shared by everything Empryo ships. The desktop app runs the same browser login (from onboarding or the keys drawer) and the same grouped model picker, and the headless CLI takes the same credentials for scripts and CI:

1empryo --headless "fix the failing test" --model llmgateway/claude-opus-5

Claude models take the gateway's Anthropic-native endpoint, so extended thinking and prompt caching survive the hop; everything else rides the OpenAI-compatible API.

Why use Empryo with LLM Gateway

  • ~190 text models — Claude, GPT, Gemini, Grok, DeepSeek, and more through one login
  • Live pricing in the picker — rates, context, and output caps come from the gateway's catalog, not a stale copy
  • Native Anthropic wire for Claude — thinking and prompt caching intact
  • Attribution built in — the dashboard shows exactly what the agent spent
  • DevPass aware — plan meters render inside Empryo, billing is handled by the gateway

Read the docs guide for routing details and troubleshooting, or Empryo's own documentation at empryo.com/docs.